Voice-activated teleprompter

A teleprompter with no speed setting.

Every conventional prompter asks you to guess how fast you are going to talk, then holds you to it. A voice-activated teleprompter removes the guess: it listens to the words you actually say and moves the script to match them.

Get it on Android Free · no account needed

Why fixed-speed scrolling never quite works

Set a scroll speed and you have committed to a single rate for the whole take. But nobody reads at one rate. You speed up through a list of features and slow down on the sentence that matters. You breathe. Someone walks past the camera.

The error compounds. Thirty seconds in you are half a line ahead of the text or half a line behind it, and now you are managing the prompter instead of talking to the lens — which is the exact thing an audience reads as this person is reading. Most people respond by nudging the speed between takes and never landing on one that fits — because the problem is not the number, it is that there is a number at all.

How it works

Three things that make it hands-free.

01

It matches words, not sound

Reacting to sound alone only tells a prompter that someone is talking. VoiceCue matches what you say against the script itself, so it knows which word you are on — which is what lets it hold position rather than guess a speed.

02

Your pace sets the pace

There is no speed slider driving anything. Speed up for a list, slow down for the line that matters, take a breath before the close — the text is where you left it either way.

03

It waits when you leave the script

Pause, ad-lib, answer a question off-camera, restart a botched sentence. The prompter stops rather than running on, and picks you up again at the right line when you return to the written words.

21 offline language models

Shipped in the mobile builds.

0 audio uploaded, on mobile

Recognition runs on the device once a language pack is installed. Nothing you say leaves it.

No account to start

Install the app, paste a script and begin. Registration is only needed for cloud sync.

When it is the wrong tool

Voice tracking is speech recognition, and speech recognition has conditions where it struggles. It is worth knowing them before you rely on it for something that matters.

  • Loud or echoing rooms. A busy café or a hard-surfaced hall degrades recognition. Fixed-speed mode is the more predictable choice there, and it is built in.
  • Dense technical or invented vocabulary. Product code names, drug names and jargon are the words a recogniser is least likely to have seen.
  • Scripts with heavy repetition. If the same sentence appears several times, any word-matching prompter has less to go on when deciding which one you are on.
  • Takes where a single slip is unacceptable. For live broadcast, a foot pedal or a human operator remains the deterministic option. Nothing about voice tracking changes that.

We have not published a cross-app accuracy benchmark yet, so we make no claim about whose voice tracking is most accurate — including ours. When that test runs, its method and raw results will be published with it. Until then, the free app is free and takes about a minute, which is a better test than any number we could quote.

Voice scrolling FAQ

Questions about voice-activated prompting.

The things people ask before they trust a prompter to keep up with them.

What is a voice-activated teleprompter?

A teleprompter that listens to you read and moves the script to match. Instead of setting a scroll speed and trying to keep up with it, you just talk. The prompter matches your spoken words against the script and keeps your current line in place — so if you slow down, it slows down, and if you pause, it waits.

How is that different from auto-scroll?

Auto-scroll moves at a constant speed you choose in advance. It has no idea where you are in the script, so any gap between your real pace and the setting compounds over the length of a take — you either run ahead of the text or fall behind it. A voice-activated prompter has no fixed speed at all. It only moves when you say the next words.

What happens if I go off script?

It stops and waits. VoiceCue pauses when your speech no longer matches the script, so ad-libbing does not drag the text away from you, and it realigns to the right place when you return to the written words. That recovery behaviour matters more in practice than raw matching accuracy — the failure people actually notice is a prompter that keeps scrolling while they improvise.

Does it work in languages other than English?

Yes. The VoiceCue mobile builds ship 20+ offline language models, including Spanish, French, German, Chinese, Japanese, Korean, Italian, Turkish, Dutch, Portuguese, Russian, Polish and Ukrainian.

Does it need an internet connection?

On mobile, no — recognition runs on the device once its language pack is installed, so nothing you say is uploaded.

Is a voice-activated teleprompter better than a foot pedal?

Not always. A pedal is deterministic — it moves exactly when you press it, in any acoustic conditions, in any language, with no recognition to go wrong. Voice tracking is better when your hands are busy or you want to forget the mechanics entirely; a pedal is better in a noisy room, on heavily technical scripts, or when you cannot risk a mis-recognition mid-take. We wrote about where each one actually wins.

Keep reading

Talk. It follows.

Paste a script and read the first paragraph out loud. That is the whole test.

Get it on Google Play